Summary: A chemical, optical, and X-ray study of a considerable number of thomsonite specimens has led to the conclusion that thomsonite and faroelite form a continuous isomorphous series. The variation in the optical properties with the chemical composition has been traced, and the relations of a number of varieties and supposed varieties of thomsonite to the typical mineral examined. The true symmetry of thomsonite is shown to be didigonal polar (C_2v). The unit cell is shown to contain 80 oxygen atoms, and (in the fully hydrated mineral) 24 water molecules; the silicon and aluminium contents are variable, but the sum of silicon and aluminium is constant and equal to 40. The sodium and calcium contents are also variable, but the sum of calcium and sodium is normally 12. Thomsonite appears to be essentially Na₄Ca₈Al₂₀Si₂₀O₈₀.24H₂O, with two isomorphous replacements, CaAl⇄NaSi and Ca⇄Na₂; the most siliceous member observed to date appears to have about 17·4 atoms Al per unit cell, while the Ca⇄Na₂ substitution leads to variations in the sum of calcium and sodium fromabout 10·3 to about 14· 8 atoms per unit cell ; but material with this sum between 11· 5 and 12· 5 is most common. There is a very marked pseudo-cell, with c-axis half that of the true cell, and, to correspond, half the atomic contents.

It has been shown that the refractive indices of thomsonite vary considerably, falling with increase in the Si/Al-ratio, and the mean refractive index ranged, in the specimens studied, from 1· 517 to 1-535. There is reason to believe that the usual diagnostic criterion for thomsonite is not invariably true, and that thomsonites may occur in which the optic axial plane is parallel, instead of perpendicular to the elongation. There is a considerable probability that such material has been observed but passed unrecognized, at least once.

Apparatus has been designed and applied to measure the vapour-pressure of thomsonite at various temperatures and degrees of dehydration, and it appears very probable that a dimorphous high-temperature form (probably to be identified with the metathomsonite of F. Rinne) exists, the transition being readily reversible. A few experiments on the absorption of vapours other than water by dehydrated thomsonite, and on the base-exchange reactions of thomsonite, have also been made. Thomsonite has been recognized from four new localities.

Chemical, optical, and X-ray studies of gonnardite have shown it to be an independent species, though very closely related to thomsonite. It is suggested that it may be identical with the high-temperature form of thomsonite (the metathomsonite of F. Rinne). Larsen's optical data for gonnardite are shown to refer to thomsonite, while there is an error of sign in the description of Lacroix. The space-group of gonnardite could not be determined, but its unit-cell, slightly larger than the pseudo-cell of thomsonite, appears to contain approximately Ca₂Na₄Al₈Si₁₂O₄₀.14H₂O.